datalab.to valid

Datalab

Datalab builds state-of-the-art document-intelligence models, available open

For agents: Document intelligence APIs and Python SDK converting PDFs, spreadsheets and images to structured output; includes OpenAPI spec and full docs in llms.txt.

document intelligencepdf conversionocr apipython sdkstructured data extractionopenapi

Their llms.txt

The beginning of the file, from our validated copy saved 2026-10-21T12:49:28.792Z. Cached copy · Source on datalab.to

# Datalab

> Datalab builds state-of-the-art document-intelligence models, available open
> source, on a managed platform, or on-prem. The models convert any document
> (PDF, DOCX, images, spreadsheets, and more) into Markdown, HTML, or JSON,
> with state-of-the-art handling of complex layouts across 90+ languages.
> Datalab accurate scores 93.5 on OmniParseBench, an open parsing benchmark,
> the top score of every parser tested, including Reducto, LlamaParse,
> Mistral OCR, Extend, Azure, and frontier LLMs. Extraction
> (accurate mode) scores 93.85 on OmniExtractBench, the top of 10 systems.
> Endpoints also cover document segmentation (splitting a combined file into
> its documents), a document agent, form filling, and track changes. Datalab maintains the open-source models
> marker, surya, chandra, and indra (75K+ combined GitHub stars) and powers document
> intelligence at leading companies. The managed platform is SOC 2 Type II
> compliant, with opt-in zero data retention, optional BAA and EU
> residency, flexible API pricing, and custom enterprise terms.

Every blog post has a markdown mirror at its URL plus ".md"
(example: https://www.datalab.to/blog/introducing-chandra.md).
The homepage, pricing, and benchmarks have markdown versions too:
https://www.datalab.to/index.md, https://www.datalab.to/pricing.md, and
https://www.datalab.to/benchmarks.md.

## Start here

- [Overview](https://www.datalab.to/index.md): what Datalab does, processors and modes, prices, benchmarks, deployment
- [Quickstart](https://documentation.datalab.to/docs/welcome/quickstart.md): get an API key and convert a document in minutes
- [API documentation index](https://documentation.datalab.to/llms.txt): every docs page, each with a markdown mirror
- [Full API docs as one file](https://documentation.datalab.to/llms-full.txt)
- [OpenAPI specification](https://www.datalab.to/openapi.json)
- [Python SDK](https://github.com/datalab-to/sdk): `pip install datalab-python-sdk`
- [Sign up](https://www.datalab.to/auth/sign_up): self-serve, free monthly usage allowance

## Product

- [Processors](https://www.datalab.to/platform/processors): Convert (lite, fast, balanced, accurate), Extract (turbo, fast, balanced, accurate), Segment, document agent, form fill, track changes
- [Pricing](https://www.datalab.to/pricing.md): per-processor rates, plans, allowances. Convert's `mode` sets its rate per 1,000 pages: `lite` $0.50 (CPU OCR; text, tables and lists, no images), `fast` and `balanced` $4, `accurate` $10
- [Benchmarks](https://www.datalab.to/benchmarks.md): OmniParseBench and olmOCR-bench results, every parser and source
- [Deployment](https://www.datalab.to/platform): managed cloud (1B+ pages in a day), VPC, on-prem and air-gapped
- Use cases: [model training data](https://www.datalab.to/use-cases/model-training), [RAG](https://www.datalab.to/use-cases/rag), [agentic workflows](https://www.datalab.to/use-cases/agentic-workflows), [compliance](https://www.datalab.to/use-cases/comp
…

Related sites

Other listed sites in AI and machine learning. See the whole category →

Machine-readable: GET /v1/resolve?q=datalab.to · GET /v1/domain/datalab.to · this page as Accept: application/json or text/markdown · MCP lookup_domain

Automatically discovered from public files. Inclusion does not imply endorsement or publisher participation.

Claim my listing on Agent Web Map for datalab.to:
1. POST https://agentwebmap.com/v1/sites/datalab.to/claims and keep claim_id and claim_secret private.
2. Publish the proof: add the line "AgentWebMap ownership: <token>" (token from step 1) to https://datalab.to/llms.txt, or use the DNS, meta-tag or well-known method the response describes.
3. POST https://agentwebmap.com/v1/sites/datalab.to/verify with {"claim_id": "...", "claim_secret": "..."}.
4. Save the api_key from the answer; it manages the listing (PATCH https://agentwebmap.com/v1/sites/datalab.to).
MCP equivalent: tools begin_ownership_claim, then verify_domain, at https://agentwebmap.com/mcp. Docs: https://agentwebmap.com/docs#verify

Corrections or removal requests: contact the directory.